taxonomists who delivered plant identifications
over the last 15 years were essential for learning the
Monteverde flora. Mike Grayum and Barry Hammel
were especially helpful in spotting problem names,
weeding out rnisidentifications and synonyms, find-
ing voucher specimens, and fine-tuning the species
lists. I also thank Bil Alverson, Mike Grayum, and
Willow Zuchowski for their careful manuscript
review.
Financial support was provided by the National
Geographic Society, the Missouri Botanical Garden,
the Jefferson L. Miller family, the National Science
Foundation, C.I.E.E., A. A. Leath, and the Portland
Audubon Society. I am grateful to the Monteverde
Cloud Forest Preserve (administered by the Tropical
Science Center), the Monteverde Conservation League,
and many private landowners who gave their permis-
sion to carry out collecting and research on their prop-
erties. The staff and associates of the Museo Nacional
de Costa Rica and the Institute Nacional de Bio-
diver si dad have always been generous with their time
and knowledge about Costa Rica's botany.TREES ON TREES
Francis E, Putzn the wind- and mist-enshrouded elfin forest in
Monteverde, trees that fall and open canopy gaps
might themselves resprout and fill the gaps they
create. Our research team measured rates of lateral en-
croachment by tree crowns bordering canopy gaps to
monitor the release of previously suppressed trees in
the understory and to compare experimentally the
contributions to canopy gap regeneration of freshly
dispersed seeds versus buried dormant seeds. Con-
trary to our expectations, more large seedlings of
canopy trees occurred in recently formed gaps than
in the adjacent understory.
Disturbed by the seemingly miraculous appear-
ance of seedlings in recently formed treefall gaps, we
looked upward and saw the source of the unac-
counted-for seedlings: juvenile trees growing as epi-
phytes on humus-laden branches in the canopy, Seed-
lings of the most common canopy trees in the elfin
forest of Monteverde that frequently grow as epi-
phytes include representatives of the Araliaceae (Den-
dropanax latilobus, Oreopanaxnubigenus, Schefflera
rodrigueziana), Melastomataceae (Blakea chlorantha),
Clusiaceae (Clusia spp.), and Rubiaceae (Cosmibuena
valerii). All were frequently encountered as epi-
phytes. Even in Monteverde's cloud forest, the canopy
is brighter than the understory. The soil that builds
up on large tree branches appears similar to the sur-
face soil on the ground. Both are mostly organic, har-
bor earthworms and other soil invertebrates, and sel-
dom dry out. Growing as an epiphyte in the cloud
forest is not much different from growing on the
ground, even for seedlings of large canopy trees. By
being perched in the canopy, they enjoy the advan-
tage of more light.
After their supporting trees fall, epiphytic tree
seedlings reorient to be upright again and send roots
down to the ground. The seedling stage is a period of
high mortality for forest trees. By first becoming es-
tablished on the branches of canopy trees, the seed-
lings circumvent the soil seed bank and avoid the
darkness of the understory. The architectural plastic-
ity needed to perform these feats of vegetative acro-
batics may seem extraordinary, but these genera (ex-
cept for Cosmibuena and Dendropanax, which are
normally trees) contain numerous hemiepiphytes
(species with a life history that includes both epi-
phytic and terrestrial phases; Putz and Holbrook
1989). Strangler figs and other hemiepiphytes are
known for their protean qualities (Dobzhansky and
Murca-Pires 1954). The "terrestrialization" process
in hemiepiphytes is normally gradual as roots grow
slowly down to the soil. In the elfin forest of Monte-
verde, however, terrestrialization often involves the
trauma of the fallen host tree and startling contortions
by the epiphytic tree seedlings carried down to the
ground (Lawton and Putz 1988).70 Plants and VegetationI